AI Thermal Detector with TA System
Model: AITD-820HT
QuixScan® Thermal Detector AITD-820HT adopts an AI chipset,
large thermal sensor array and latest algorithm which provides
cutting-edge face recognition and body temperature detection, with
the ability to detect even with a mask on.
Check in your staff..
The AITD-820HT can check temperature, acts in access control, unlock
doors and at the same time assist in checking in your staff without
contact.
High Accuracy
±0.5 ° F forehead temperature detection
99.9% Face recognition: the best solution for access control and time attendance system
Fast Detection
(0.3s Temperature + face ID)
Wide detection Range
0.3 m ~ 1.2 m forehead body temperature detection
0.5 m~ 2 m face recognition
Wear mask recognition
Support face recognition with face mask, recognition rate >95%
Supports 1: 1 or 1: N face comparison
Attribute recognition
Supports detection of race, hard hat, skin color, etc.
Super backlight suppression
Support face exposure, automatic adjustment according to strong light
Support nighttime 3D noise reduction, provide best quality in darkness
Variety of door access control
Support face recognition or door access control
Customizable multi-language
English, Spanish, Italian, Russian, Japanese, Korean.
Hardware Specifications
|
Model |
AITD-820HT |
|
Item |
Specification |
|
Sensor Type |
Uncooled infrared focal plane array |
|
Infrared Thermography camera |
120*90 |
|
Thermal measurement distance range |
1 FT. - 4 FT. |
|
Thermal measurement accuracy |
±0.5ºF |
|
Hardware |
|
|
Processer |
AI chipset ARM Cortex-A7 quad-core, 4 x high performance CEVA DSP |
|
OS |
Linux |
|
Storage |
8GB EMMC+1GB DDR4 |
|
LCD |
7 inch, 1024*600 resolution |
|
Camera |
|
|
Sensor |
Dual 2Mp, low lux sensor |
|
WDR |
≥120dB |
|
FoV (Diagonal/horizontal/ vertical) |
73°/65°/40° |
|
Iris |
F2.0 |
|
Focus length |
4.3mm |
|
Performance |
|
|
Anti-spoofing |
Anti-electronic screen, video, mask hacker attack |
|
Face Detection range |
0.5m~2m, lens adjustable |
|
Face library |
50,000 (default) |
|
Attendance Record |
50,000 |
|
Face angle |
horizontal 30 degree, vertical 30 degree |
|
Interface |
|
|
Access control |
Push out |
|
Power supply |
DC12V , 2A ( 9 -16V DC) |
|
Power consumption |
Less than 20W |
|
Ethernet |
1 pcs RJ45 10M / 100M Ethernet |
|
Wifi |
2.4G Wi-Fi |
|
USB |
1 pcs USB 2.0 |
|
General |
|
|
Ambient Working Temperature |
-40° F - 100° F |
|
Dimensions |
4.75"X 8.9" X 1.3" |
Algorithm Parameters
|
Application |
Items |
Description |
|
Face recognition |
Face recognition |
Local recognition: 50,000 (default ) |
|
Algorithm accuracy |
99.9%@0.01% （The accuracy rate is 99% under the 10,000th false alarm rate） 95%@0.0001% （The accuracy rate is 95% under the one-millionth false alarm rate） |
|
|
Angle requirements |
Yaw ±30°，roll ±30°，pitch ±15° |
|
|
Face minimum pixels |
Face size above 80 * 80 pixels |
|
|
Anti-spoofing |
Angle requirements |
Yaw ±30°，roll ±30°，pitch ±15° |
|
Face pixel requirements |
Interpapillary distance 50 pixels and above |
|
|
Algorithm accuracy |
98%@0.01% （Real pass rate@Dummy pass rate） |
|
|
Anti-spoofing category |
Pictures, videos, coated paper, fake face models, etc. |
|
|
Face detection and tracking |
Number of people captured |
More than 30 people / frame |
|
Snap frame rate |
25fps |
|
|
Minimum face size |
30*30 |
|
|
Adaptation angle |
Yaw ±45°，roll ±45°，pitch ±30° |
|
|
Face attribute |
Item |
Gender, age, glasses, facial expression (angry, calm, happy, sad, surprised), mask, ethnicity, beard, ethnicity, hat detection, helmet detection |
|
Algorithm accuracy |
Age accuracy (± 5 years old): more than 85%; Gender accuracy: more than 96%; |
